When it's critical that she actually know what the weather will be like, she asks me, and I look it up from NOAA, which is extremely accurate, as long as you look up the forecast for an actual whether station, instead of using a location between stations.
The biggest challenge in whether forecasts isn't the forecast itself; it's in communicating them. They're especially bad at differentiating changes in probability over time vs location.
That's great, but it certainly doesn't address "why you hate your weather app." And "sometimes the forecast isn't perfect" is not a reason I've ever had. Poor behavior, ad bombardment, poor design, bad visibility, and more are all reasons I've had in the past, but never "the forecast isn't perfect."
I got fed up with several things in Apple Weather but they appeared to be lack of care and attention to design, rather than underlying data sources. One was the lock screen widget not showing current conditions and showing "severe weather" instead. The severe weather was air quality alerts. Less than useless. Another was NWS alerts coming up with generic data rather than the actual information, so I often got recommendations along the line of "execute a pre-planned activity". Useless again.
Perhaps that is because I work in renewable energy and am very familiar with the weather data models and sources and experienced with the variability, but surely we've all experienced weather's unpredictability before.
By the way, I like my current weather app, Breezy[1], which uses a number of sources including Open Meteo[2].
